Big ideas

The conceptual core of tractor: what structured concurrency (SC) means once your “tasks” are whole processes, and how the runtime is layered to deliver that without (much) magic.

If you only read one page in these docs make it Structured concurrency, across processes; if you read two, follow it with Anatomy of the runtime.

• Structured concurrency, across processes
  • SC in one breath
  • The leap: process-shaped tasks
  • Causality: no process outlives its parent
  • Errors always propagate (yes, across the wire)
  • Cancellation is a request, supervision is the rule
  • Hold up, is this an “actor model”?
  • Why processes at all?
  • It’s just trio
• Anatomy of the runtime
  • The layer cake
  • One actor, one process, one trio.run()
  • IPC: channels, transports, addresses
  • Spawn backends
  • The registrar
  • Runtime env vars
  • Logging
  • Where to next?